Cathie Wood’s ARK sells AMD stock, buys Tempus AI By Investing.com

ARK disclosed a portfolio shift on May 7, 2026, led by a $8.29M purchase of 155,045 Tempus AI shares across ARKK and ARKG, following a similar buy the prior day. The firm also added 102,247 Intellia Therapeutics shares for $1.42M, while trimming AMD by 8,045 shares ($3.39M), Teradyne by 9,210 shares ($3.52M), and smaller positions in CareDx, Twist Bioscience, Veracyte, and Adaptive Biotechnologies. The trades indicate continued rotation toward AI/biotech innovation names and away from some established semiconductor holdings.

Analysis

The signal is less about the individual biotech adds than the factor rotation underneath: capital is being pulled from mature, high-beta semiconductor automation into earlier-stage therapeutic/platform names where data readouts can reprice the stock faster. That usually happens when a growth allocator expects dispersion to widen, not narrow — i.e., the market is rewarding idiosyncratic catalysts over multiple expansion. In that setup, names with crowded AI/semicap ownership can underperform even if the macro tape is stable, because incremental buyer support is weaker and any disappointment gets amplified.

For AMD and TER, the risk is not just fund-flow pressure; it is that these names become the easiest liquidity source in a de-risking sequence if AI spend visibility slips even modestly. If hyperscaler capex commentary softens over the next 1-2 earnings cycles, the downside can be swift because the stocks are trading with a lot of forward optimism embedded. By contrast, TEM and NTLA are benefiting from a different kind of optionality: investors are paying for clinical or product milestones, so the payoff profile is asymmetric but binary, with outcomes likely to be decided over the next 3-9 months rather than on quarterly beats.

The contrarian angle is that the selling in the semis may be overshooting the fundamental signal. If this is a pure portfolio rebalance rather than an information edge, then AMD and TER can mean-revert quickly once the market re-focuses on AI infrastructure demand, while the biotech buys may be vulnerable to dilution and trial-timing risk that is easy to underestimate. The more interesting expression is to fade the crowded narrative rather than the stock: leadership in AI hardware can reassert if order revisions stabilize, but the path is choppy until the next catalyst confirms that capex is intact.

Key Decisions for Investors

• Short-term pair: long TEM / short AMD into the next 2-6 weeks if you want to express catalyst dispersion; target a 1.5-2.0x payoff if TEM gets additional accumulation while AMD loses marginal flow support. Use a tight stop if AMD reclaims relative strength on AI capex headlines.
• Take a tactical short or buy puts on TER over the next 1-2 months only if channel checks show slowing equipment orders; otherwise this is a flow-driven trade with limited edge. Risk/reward favors defined-risk puts rather than outright short stock.
• Accumulate NTLA on weakness for a 3-6 month horizon, but size it as an event-driven lottery ticket: the upside can be substantial on positive clinical/data updates, while the main risk is dilution and timing slippage. Prefer call spreads over common if implied vol is elevated.
• Fade the semicap de-risking with a relative-value long AMD vs. a broad software/growth basket if AI infrastructure sentiment remains intact. The trade works best if you expect the current selling to be a transient allocator rotation rather than a thesis change.
• Avoid chasing CDNA/TWST/VCYT/ADPT on the long side until there is proof of sustained institutional sponsorship; these names can underperform for weeks after flow reversals because liquidity is thin and the market discounts them faster than fundamentals update.